Similarity for Audio : from music to environmental sounds

N.B. : La partie sonore sera en français, la partie visuelle sera en anglais. Autrement dit, Mathieu parlera français même si ses diapos sont en anglais.

Résumé :

Similarity-based tasks are meaningful when considering scientific problems related to the representation of sound signals. The first part of the talk will be dedicated to a critical review of the use of this notion when considering musical signals. This will lead us to consider environmental sounds. In particular, it might be easier with this kind of sound to demonstrate the potential benefits of specific sound representations as well as relating them to knowledge from neuroscience and psycho-perception.

Notre présentateur :

Mathieu Lagrange (IRCAM, Paris & IRCCYN, Nantes) is a CNRS researcher at IRCCYN. His main interest is the modeling of audio signals. His expertise covers numerous domains such as computer science, signal processing and machine learning. He also collaborates with renown experts in the field of neuroscience psycho perception. After a Phd in Computer Science at Orange Labs, he carried out several post doctoral fellowships in Canada (University of Victoria, BC and McGill University, QC) and in France (Telecom ParisTech).